We draw on predictive mind theory—an emerging paradigm shift in the cognitive and computational neurosciences—to characterize affordance perception and enactment in the context of digital technologies as a generative process. Key to our argument is that affordance perception and enactment are characterized by a forward flow from the agent to the object. Human and autonomous agents continually generate and project affordance predictions about digital technologies that are rooted in their previous knowledge. They continuously adjust these affordance predictions as they are exposed to unexpected prediction errors. We identify two central mechanisms: reactive and proactive generative processes. Reactive generative processes describe situations where agents generate new affordances when they accidentally encounter prediction errors. Proactive generative processes describe situations where agents embrace prediction errors and explore new action opportunities. Our theorizing highlights how agents and digital technologies are continually changing in a dynamic learning process and provides a conceptual foundation for studying affordance innovation through human and autonomous agents.